The Test Bench › Opening range breakout › Study 006
We backtested the gap and go strategy on 149 large US stocks, and checked whether waiting for the breakout actually helps.
STUDY 006Gap and go is a day-trading strategy for stocks that gap up, meaning they open higher than they closed the day before. The idea is that the rise continues. A common way to time the entry is the opening range breakout: buying when the price climbs above the high of the first minutes of trading.
We backtested it on 149 large US stocks over six months. On days when a stock opened more than 0.44% higher, the average trade made +0.033% after trading costs. On days that opened flat or lower, the average trade lost money.
Waiting for the breakout did not help. Buying as soon as the first 15 minutes of trading ended made +0.055% a trade. Waiting for the breakout made +0.054%. Both were held until the end of the day.
The gain was small, and a few big days did much of the work. Without the stocks that opened more than 5% higher, it was +0.017% a trade.
This was a second look at the data from our previous opening range breakout study, not a fresh test. It does not show that gap and go will be profitable in future.
These are historical analyses. They are neither advice nor forecasts. A backtest describes what a rule would have done over a past window. It does not say what the rule will do next.
Study 005 tested one 15-minute opening range breakout rule on 149 large US stocks and found a negative average trade after costs. Split by how each day opened, gap-up days were the only group with a positive average in both parts of its six-month window. This study examines that split on the same data.
It asks three questions:
Every trade here comes from the same candles, stocks and six months as Study 005. The splits were chosen after Study 005's results were known. In total, 17 ways of splitting the trades were examined, on 12 versions of the rule and 5 purchases without a breakout. The results are exploratory: they describe the data Study 005 used and are not an independent test of gap and go.
Every signal is read at the close of a 5-minute candle. Every order, including each partial sale, target and end-of-day exit, fills at the open of the next candle.
| Part | Definition |
|---|---|
| Opening gap | The day's first 5-minute open against the previous day's last 5-minute close on IEX. That close is the last IEX trade of the session, which can differ from the official closing price. |
| Gap up, flat, gap down | A gap up is above +0.444%. A gap down is below -0.281%. Anything between is flat. These cut-points divide Study 005's first-period trades into three equal groups. They come from this data and are not standard definitions of a gap. |
| Opening volume | The IEX volume of the opening range against the average for the same minutes over the previous 20 trading days (at least 10). It is known when the range completes, before any entry. |
| Volatility | The 14-day average true range divided by the close, from daily candles up to the previous day. The thirds are cut across every stock and day in the window, so they describe the sample. |
| Opening range | High and low of the first 15, 30, 45 or 60 minutes from 09:30 New York time: 3, 6, 9 or 12 five-minute candles. |
| Entry | A candle's low is above the range high. Buy at the next candle's open, at most once per stock per day. |
| Study 005's exits | Sell 70% after a close above the range high plus one range. Sell the rest after a close below the entry price, a close above the range high plus 2.5 ranges, a close below the range low, or at the end of the day. |
| Partial sale, then hold | Sell 70% after a close above the range high plus one range. Hold the rest to the end of the day. |
| Hold to the close | Hold the whole position to the end of the day. |
| End of day | Sell anything still open at the open of the day's last candle, 15:55. |
| No breakout | Buy at the open of the first candle after the range completes (09:45 for a 15-minute range), or at 09:35, and hold to 15:55. It buys on every gap-up day, including days with no breakout. Same fills and costs as the breakout. |
| Trading cost | 0.02% of the value of each purchase and each sale (assumed), about 0.04% a round trip. Also shown with no cost. Spreads, slippage and fees vary in practice; Study 005 sets out the cost assumption in full. |
| Portfolio | $10,000 split equally across the 149 stocks. Each share compounds its own stock's trades and stays in cash, earning nothing, between them. |
Unless stated otherwise, results use Study 005's rule: the 15-minute range with Study 005's exits.
| How the day opened | Trades | First period | Second period | Whole window | Whole window, no cost |
|---|---|---|---|---|---|
| Gap down | 3,248 | -0.105% | -0.026% | -0.079% | -0.039% |
| Flat | 3,140 | -0.108% | -0.115% | -0.110% | -0.071% |
| Gap up | 3,029 | +0.039% | +0.018% | +0.033% | +0.073% |
Gap-up days were the only group above zero, in both periods. Both periods were part of the exploratory work, so this is not an out-of-sample test.
Traded only on gap-up days, $10,000 split across the 149 stocks finished at $10,070 after costs (+0.7%). The same rule on every day finished at $9,674. The gap-up version made 3,029 trades against 9,417, so its money sat in cash more of the time.
| Opening gap | Trades (first · second period) | First period | Second period | Whole window |
|---|---|---|---|---|
| Below -5% | 86 (68 · 18) | -0.148% | -0.061% | -0.130% |
| -5 to -3% | 167 (132 · 35) | -0.140% | +0.376% | -0.032% |
| -3 to -1% | 1,119 (747 · 372) | -0.214% | +0.001% | -0.142% |
| -1 to -0.28% | 1,876 (1,233 · 643) | -0.033% | -0.063% | -0.043% |
| Flat | 3,140 (2,179 · 961) | -0.108% | -0.115% | -0.110% |
| +0.44 to +1% | 1,372 (935 · 437) | +0.050% | -0.053% | +0.017% |
| +1 to +3% | 1,333 (994 · 339) | +0.025% | -0.010% | +0.016% |
| +3 to +5% | 204 (157 · 47) | -0.060% | +0.260% | +0.014% |
| Above +5% | 120 (94 · 26) | +0.244% | +1.140% | +0.438% |
Excluding gaps above +5% reduced the gap-up average from +0.033% to +0.017%, over 2,909 trades. The 120 trades on those gaps averaged +0.438%, the highest of any band; 26 of them fell in the second period. Gaps between +0.444% and +3% were not positive in both periods.
Each cell is the average trade after costs over the whole window, with the number of trades.
| How the day opened | Opening volume below its average | 1 to 2 times its average | Above 2 times its average |
|---|---|---|---|
| Gap down | -0.140% (1,872) | +0.023% (819) | -0.280% (256) |
| Flat | -0.126% (1,996) | -0.106% (683) | -0.112% (189) |
| Gap up | -0.001% (1,656) | +0.005% (888) | +0.163% (346) |
On gap-up days with opening volume above twice its usual level, the average trade was +0.163% over 346 trades from 137 stocks, higher than at lower volume. On gap-down days with the same volume it was -0.280%, the lowest of any group. Flat days were negative at every level of volume. Volume is compared with each stock's own history. This is one of the exploratory splits, not a tested filter.
Average trade on gap-up days after costs, first period · second period.
| Opening range | Study 005's exits | Partial sale, then hold | Hold to the close |
|---|---|---|---|
| 15 minutes | +0.039% · +0.018% | +0.050% · +0.026% | +0.051% · +0.062% |
| 30 minutes | +0.060% · +0.032% | +0.062% · +0.043% | +0.051% · +0.058% |
| 45 minutes | +0.044% · +0.061% | +0.046% · +0.053% | +0.031% · +0.056% |
| 60 minutes | +0.025% · +0.052% | +0.027% · +0.041% | +0.018% · +0.036% |
On gap-up days, all 12 combinations of range length and exit averaged above zero in both periods, between +0.022% and +0.057% over the whole window. On flat days, all 12 were negative. Across every day, all 12 were negative after costs. The sign of the gap-up and flat difference was the same across the tested variants in this sample.
The breakout waits for a candle to trade above the range. The comparison buys every stock on every gap-up day as soon as the range is complete, and holds to the same 15:55 exit, at the same cost.
| Opening range | Breakout, held to the close | No breakout, bought at range end | Trades (breakout · no breakout) |
|---|---|---|---|
| 15 minutes | +0.054% | +0.055% | 3,029 · 5,944 |
| 30 minutes | +0.053% | +0.052% | 2,655 · 5,944 |
| 45 minutes | +0.037% | +0.082% | 2,425 · 5,944 |
| 60 minutes | +0.022% | +0.035% | 2,235 · 5,944 |
| No breakout, bought at 09:35 | +0.062% | 5,944 |
At 15 and 30 minutes the two averages were within 0.005 percentage points of each other. At 45 and 60 minutes the purchase without a breakout averaged more. Buying at 09:35 averaged +0.062%. In these comparisons, waiting for the breakout did not raise the average trade above buying at the same time without it.
The comparison is descriptive. The purchase without a breakout trades on all 5,944 gap-up stock-days, the 15-minute breakout on the 3,029 with a breakout, so the two groups of days differ. On flat days, the 15-minute breakout held to the close averaged -0.136% and the purchase at 09:45 -0.061%.
On gap-up days, 79 of the 149 stocks finished with a positive total under Study 005's rule. The trade returns of the five strongest, MRNA, CRWD, OKTA, PANW and PLTR, added up to 87% of the net sum of all gap-up trade returns, winners and losers together. The positive average was not spread evenly across the stocks.
Split by volatility, the most volatile third of gap-up trades averaged +0.121%, the middle third +0.006% and the least volatile third -0.064%. Only 89 of the 149 stocks had any trade in the most volatile third.
Each stock's gap-up days were also split at that stock's own median volatility over the window (149 stocks with at least 6 gap-up trades). Its more volatile days averaged -0.012% and its calmer days +0.076%. The pooled volatility pattern did not survive this within-stock comparison. It appears to reflect differences between stocks rather than better results on a stock's more volatile days.
Study 005's rule lost money on average. Split by how the day opened, gap-up days were the only group with a positive average, +0.033% a trade after costs, and the result was small: $10,000 traded only on gap-up days ended at $10,070. It rested on a few stocks, and without gaps above +5% the average fell to +0.017%.
On gap-up days, buying without a breakout returned about the same as the breakout, or more. The analysis points to a question about gap-up days, not to an edge in the breakout. Testing that question needs data none of these choices touched.
The opening range, the breakout entry, the partial sale, the exits and the trading cost used here are blocks on GU Analyser's canvas. You can rebuild Study 005's rule, change the range length or the exit, and test it on the stocks you trade.
Using this research. The analysis, tables and figures on this page are GU Analyser Ltd’s own work, published under a Creative Commons BY-NC 4.0 licence. You are welcome to quote it, cite it and build on it, with credit and a link. Suggested credit: GU Analyser, The Test Bench, Study 006, linking to guanalyser.com/test-bench/gap-and-go/. For commercial use, ask us at hello@guanalyser.com.
Rebuild the test, or test your own gap setup
On the strategy canvas: the opening range of the first 5-minute candles, a buy after a candle whose low is above ORB High, a partial sale at a first target and an exit by the end of the day. Then change the range length, the exit or the cost, and run it on the stocks you trade.
Create a free account and build your own strategies → or read how the Opening Range is calculated →The method: how a backtest is built